Our take
Quietno commits in the last 90 days and no release in a year- Activity
- 19
- Community
- 39
- Issues
- 21
- Pull requests
- 20
+0 stars in the last 30 days
A focused option for Claude Code users who want connected, evolving project memory, though its quiet recent development makes it worth evaluating before relying on it for critical workflows.
Good fit if
- You use Claude Code and want an MCP-based memory tool with semantic and graph search.
- You can configure an LLM backend and want project-specific or shared ChromaDB storage.
- You value a permissively licensed MIT project and can assess its current behavior against your needs.
Look elsewhere if
- You need a project with recent commits or releases as evidence of ongoing maintenance; this repository has had neither in the measured period.
- You need documented support for agents beyond Claude Code today, rather than planned compatibility.
- Your workflow cannot accommodate an LLM backend and ChromaDB-based storage.
All health signals
| Last commit | 2026-01-17 (9 months ago) |
|---|---|
| Commits, last 90 days | 0 |
| Releases, last 12 months | 0 |
| Contributors | 4 (top contributor: 61% of commits) |
| Issues closed, last 90 days | 0 |
| Pull requests merged, last 90 days | 0 |
| Project age | 10 months |
Checked on 2026-10-28 with the GitHub API.
Project history
Daily GitHub snapshots since Oct 7, 2026. History covers stars and forks; health signals are not tracked historically.
Stars
35Forks
3Change since Nov 17, 2026: 0 stars·0 forks
Overview
A-MEM MCP brings the A-MEM self-evolving memory approach to coding agents, using ChromaDB to store memories and their relationships. Instead of treating notes as isolated vector-search results, it builds a Zettelkasten-style graph that can connect and update related knowledge as new memories are added.
It is an MCP server currently tested with Claude Code. It is aimed at developers who want an agent to retain project context across sessions, such as architectural decisions or implementation details. The README says support for other MCP-compatible agents is planned.
Key Features
- Extracts keywords, context, and tags from new memories using an LLM.
- Finds semantically similar memories and can strengthen links or update related entries.
- Combines semantic search with graph traversal to discover connected knowledge.
- Offers a peek-and-drill workflow: inspect lightweight metadata first, then read full notes as needed.
- Provides MCP tools for adding, searching, reading, updating, and deleting memories, plus checking asynchronous task status.
- Supports time-filtered search and bulk reads.
- Offers a Python API in addition to the MCP server.
Use Cases
- A Claude Code user can save project conventions and retrieve them during later sessions.
- A coding agent can look up related past decisions when investigating an unfamiliar feature or bug.
- Developers can build a searchable record of implementation details that gains links as the project evolves.
- Applications using Python can access the memory system directly without integrating through MCP.
What you need
Detected in the repository
- Python >=3.8 (from pyproject.toml)
- A test suite
License in plain words
MITpermissive
- Commercial use: yes
- Modify and redistribute: yes
- You must keep: the copyright and license notice
- Share your changes: no
A summary, not legal advice: the LICENSE file is what applies.
Getting Started
Install the package:
pip install a-mem
For Claude Code configuration and environment variables, see the repository README.
Alternatives
- open-index: Open Index uses a general-purpose context graph with SQLite or OpenSearch, rather than A-MEM’s evolving project memory stored in ChromaDB.
- agentmemory: agentmemory captures and shares coding-agent memories across sessions and tools, using iii-engine rather than A-MEM’s ChromaDB knowledge graph.
- Agent-Memory: Agent-Memory focuses on automatic cross-session context capture and recall for coding agents, rather than A-MEM’s evolving linked-memory approach.
- Memary: Memary builds interaction-derived entity graphs for autonomous agents with LlamaIndex and a ReAct agent, rather than project-focused coding-agent memory.
| Project | Language | License | Stars | Status |
|---|---|---|---|---|
| a-mem-mcp | Python | MIT | 35 | Quiet |
| open-index | Python | MIT | 112 | New |
| agentmemory | TypeScript | Apache-2.0 | 29.1k | Active |
| Agent-Memory | – | Apache-2.0 | 9 | Quiet |
| Memary | Jupyter Notebook | MIT | 2.7k | Inactive |
Considerations
The project is marked Quiet: there have been no commits in the last 90 days and no release in over a year. Four contributors are recorded, and activity is not concentrated enough to identify dependence on one person as a specific risk. The repository has tests, but no CI is reported; the measured data also shows no recent issue closures or merged pull requests. Python 3.8 or newer is required. The README documents configurable LLM backends and ChromaDB storage, so teams should account for those dependencies and decide whether project-local or shared memory fits their data boundaries. The README identifies Claude Code as the tested agent and describes broader MCP-agent support as planned.
Found this useful?
Share it with someone who would like a-mem-mcp.